VISNIR: Machine Learning Model for Copper Concentrate Classification

Franco Rivas Paredes developed a VISNIR model for classifying copper concentrates based on reflectance spectra. The model is intended for use with a specific reflectance dataset published on figshare in 2023. Matlab® software is required to run the provided models.

Use Cases

• Classifying copper concentrate mineralogy based on VISNIR-SWIR reflectance spectra.
• Validating spectral models against QEMSCAN mineralogical characterization data.
• Developing predictive models for industrial mineral processing quality control.

Strengths

• Model is explicitly linked to a published, peer-reviewed reflectance dataset.
• Associated mineralogical ground truth data is available from a separate QEMSCAN report.

Limitations

• Column-level documentation for the model inputs and outputs is absent.
• Row count and dataset scale are unknown, limiting suitability assessment.
• The model requires proprietary Matlab® software to run, which may limit accessibility.

Provenance

Source

Franco Rivas Paredes, via paperswithcode.

Collection Method

Model developed for use with a specific reflectance dataset; creation method not detailed.